Send email for updates

Effectiveness bank home page. Opens new windowHot topic

Below is one of our selection of Hot Topics, important issues which sometimes generate heated debate over the facts or their interpretation. Click the GO button or the Blue title to trigger a customised search for relevant Findings logo documents. Links to other documents. Hover over for notes. Click to highlight passage referred to. Unfold extra text Unfold supplementary text

Copy title and link | Comment/query | Other hot topics |

How many drinkers should be in treatment?

How well is the UK doing at getting people who need this help into treatment for their drinking problems? It’s a statistic which matters, because the more of the in-need population we treat, the smaller the alcohol-dependent population and the less the related harm. We can get an idea of how much better the UK could feasibly be doing by comparing countries within the UK, where Scotland seems to be doing much better than England (at least three times better) at meeting treatment need (1 2). An estimate for Leeds in northern England is that trebling access to treatment to the level in Scotland would over five years cut the alcohol-dependent population by just under a fifth and save a further 65 lives. In 2004, UK-wide there would have been an estimated 794 fewer deaths had one in five dependent drinkers been treated with medications versus if none had. Numbers avoiding illness would have been considerably greater.

The aim is to present a ball-park indication not of absolute numbers and proportions needing treatment, but of the degree to which these figures alter under different criteria for what counts as ‘needing treatment’. Our analysis focuses on England and the year 2014 because this is where the data and estimates are most complete. Most pertinently, this was the year and England was the country addressed by the key source for these estimates panel.

As we’ll see, depending on where you draw the line, England’s performance in ensuring needy drinkers enter treatment can look anywhere from a poor 7.5% to an excellent 43% or even more. Line-drawing is a matter of judgement, and perhaps too of motivation – of how you want to portray performance, and in turn whether you want to advocate for more services, or to reassure that need can already largely be met. But as a basis for these judgements, we should be as clear as possible about the relevant facts and uncertainties.

What proportion is the treatment caseload of drinkers in need of treatment?

There was a reason why the heading above did not read, “What proportion of drinkers in need of treatment are actually in treatment?” Though this is the crunch question, it cannot be answered. We know the numbers in treatment and can estimate the in-need population, but cannot know whether every patient has come from that in-need population. Despite not qualifying according to the criteria used to assess need, some may have found their way or been directed into treatment, while others may still be in treatment long after they no longer score as in-need. This gap in the available data means all we can say is that the numbers in treatment are equivalent to a certain proportion of the in-need population, implying that up to (but not necessarily all) that proportion are in treatment. For example, at their intake assessments at most 57% of the adults in treatment for their drinking problems in 2013/14 in England had said they had been drinking 16 UK units or more a day. It means many must have been drinking below the 15 units a day which NICE says indicates a need for treatment.

Key source

The key source for estimates of alcohol treatment need in England is a report commissioned by Public Health England from researchers at the University of Sheffield and King’s College, London. Among other things, it was intended to establish the extent of the need for specialist treatment of problem drinking in England in 2014/15 based on the estimated alcohol-dependent population in 2014. In turn it drew its data on the alcohol-dependent population from a survey of the mental health of the English population conducted in 2014, the Adult Psychiatric Morbidity Survey 2014.

In the calculations which follow, the top element of the needs-met/total need fraction is stable – the number recorded as being in treatment or the treatment ‘caseload’. In England, 114,920 adults aged 18 or over were recorded as having been in specialist treatment primarily for the treatment of their drinking problems at some time during the year 2013/14, the last time the figures were presented in this way. For subsequent years we know the number of patients whose substance use problems included alcohol and the far lesser number with alcohol as their sole substance use problem. Somewhere between these numbers is the number primarily treated for their drinking problems – of greatest relevance, because it is the alcohol-treatment caseload as opposed to the caseload who drink but are primarily being treated for a drug problem. If the relative position of that number in the range stays as it was the year before, in 2014/15 some 111,573 patients would have been in specialist treatment primarily for the treatment of their drinking problems – our best estimate for the top element of the fraction in the year 2014.

In contrast, the bottom element of the fraction can vary widely, depending on how many adults meet different criteria for being in need of treatment. The resulting fractions express the treatment caseload as a proportion of the variously defined ‘in-need’ population. As we’ll see, variation in how that population is defined means the fraction can vary from worryingly small to reassuringly high figure. Below are some answers to the question posed for this section (“What proportion is the treatment caseload of drinkers in need of treatment?”), drawing largely on the key source’s quite sophisticated estimates of treatment need.

The proportion of drinkers in England in treatment of those in need of treatment under different assumptions

Is it 7.5%? The Adult Psychiatric Morbidity Survey 2014 found that 3.1% of people aged 16 or over scored at least 16 on the AUDIT screening questionnaire, indicative of harm from their drinking and/or least mild dependence. Using this survey, our key source estimated that in 2014/15, 1,491,458 – about one-and-a-half million – people aged 18 or over or were drinking harmfully or exhibited some degree of dependence, 3.5% of the same-age population. Combining this figure with the treatment caseload in 2014/15 yields an estimate that the equivalent of 7.5% of England’s harmful or at least mildly dependent drinkers had been in specialist treatment for their drinking problems during 2014.

A different kind of AUDIT-based estimate has been derived from the Alcohol Toolkit Study (of which more below), which mounted national surveys in England during 2014 to 2017. Numbers accessing treatment were not based on the national treatment caseload, but on actual reports from survey respondents. About half who scored 16 or more on the AUDIT had attempted to cut back and reported on the services (if any) they had accessed at their last attempt; just under 9% said they had attended a specialist alcohol clinic or centre, equating to about 4% of all respondents who scored 16+. When the AUDIT score was 20 or more (indicative of probable dependence) the proportions were about 16% and 8% respectively.

Is it 11%? Using the same key source we can narrow the one-and-a-half million ‘in-need’ estimate down to the roughly a million (precise estimate, 1,023,587) adults who in addition to the AUDIT 16+ criterion, also scored as at least mildly dependent on alcohol on the Severity of Alcohol Dependence Questionnaire (SADQ). As its name suggests, unlike AUDIT the SADQ is expressly intended to assess severity of dependence; the version of the questionnaire used in the relevant surveys is the ‘community’ version adapted for non-clinical samples such as the general population. According to NICE, Britain’s official authority on health interventions, drinkers who mee t these criteria should be considered for less intensive forms of treatment, including care coordination and psychological interventions, and occasionally anti-relapse medication or aftercare. On this basis, the estimated treatment caseload in 2014/15 was equivalent to 11% of the at least mildly dependent drinkers whose drinking was causing harm and who qualified for some form of treatment. One concern over this estimate is that by design, the Severity of Alcohol Dependence Questionnaire was not based on the clinical criteria used to diagnosis dependence.

Is it 43%? Putting that concern to one side, results from the SADQ gathered by the Adult Psychiatric Morbidity Survey 2014 can be used to narrow down further to the cohort who perhaps really ought to be in treatment. In 2011 NICE calculated that in England 260,000 adults were not just (based on an AUDIT score of at least 16) harmfully drinking and perhaps mildly dependent, but “moderately” dependent, indicated by their also scoring at least 16 on the Severity of Alcohol Dependence Questionnaire. Our key source updated that figure to 257,626 for 2014/15. Accepting this as the in-need population suggests that in 2014 the treatment caseload was equivalent to 43% of adults whose harmful and dependent drinking ‘really’ justifies specialist help. What kind of help can be extrapolated from NICE’s treatment recommendations. Though not identifying this precise group, we can assume they warrant at least the response specified for a severity step down (the possibly mildly dependent, harmful drinkers of the previous calculations) and possibly that specified for drinkers a step up, whose AUDIT scores of 20 or more indicate not just harmful drinking, but probable dependence. For these drinkers the recommendations extend to comprehensive assessment, managed community withdrawal, and definite consideration of anti-relapse medication and aftercare.

Is it 19%? But the key source itself used a wider bracket to estimate potential treatment need. To the total described in the previous paragraph, it added 337,505 people who though they may have scored as mildly dependent on the Severity of Alcohol Dependence Questionnaire (score 4–15), registered an AUDIT score of at least 20, indicative of probable dependence. Put another way, this method includes anyone who scores as a high-risk and probably dependent drinker on AUDIT (20 or more) unless that ‘probability’ is flatly contradicted by a SADQ score variously described as indicative of no dependence at all or a severity not even rating the term ‘mild’. It then adds in anyone with a lesser AUDIT score of at least 16 (indicative of harm from their drinking and possibly mild dependence) as long as this is bolstered by a SADQ score of 16 or more, indicative of at least moderately severe dependence. The result was a total of 595,131 “adults in England with alcohol dependence”, a figure accepted by Public Health England as the population who “may need treatment”. The implication is that the 2014/15 alcohol treatment caseload was equivalent to 19% of the alcohol-dependent adult population.

Based on clinical guidelines developed for NICE, this formula for estimating treatment need had been constructed for the Department of Health by the team behind the key source. The aim was to exclude drinkers who despite a high risk to their health scored as non-dependent – suggesting a brief intervention would suffice in lieu of fully-fledged treatment – but to include those who though at slightly lower risk, were dependent enough to mean they would not remit, even after an extended brief intervention. For these drinkers, only treatment had a reasonable chance of breaking their dependence. A flaw in this method is that brief interventions may actually work best for more dependent and heavier drinkers. Another is that the evidence that brief interventions would suffice for non- or less-dependent drinkers is weak. Nevertheless, their problems may be considered insufficient to warrant treatment or to make treatment acceptable to the drinker, and their risky drinking is likely to remit without formal intervention (1 2 3).

Could it be over half? Now we have a range from treatment capturing numbers equivalent to well under 1 in 10 (7.5%) harmful and perhaps at least mildly dependent drinkers, to its capturing approaching half (43%) those also at least moderately dependent. The lower figure can be justified as a proportion of all drinkers who might gain harm-reduction benefits from treatment, the higher as based on an estimate closer to those who due to more severe dependence really do need treatment to prevent continued harm. In between is the 19% figure, intended to be the best estimate of treatment need, despite 57% of the supposedly ‘in-need’ population scoring as only mildly dependent.

The 43% figure gains support from findings indicating that most dependent drinkers do not need treatment to overcome their dependence. In the USA, three-quarters remit without treatment and just 10% are clearly both in need of treatment and most often access it. In the UK NICE has also appeared to elevate the treatment-need line closer to (and even above) the severity level which yielded the 43% estimate. Rather than 16 on both, their criteria were a score of at least 20 on AUDIT and 16 on the Severity of Alcohol Dependence Questionnaire, implying that in 2014 England was treating the equivalent of more than 43% of the in-need population. Just what that proportion is according to these NICE criteria is unclear, but it must be around or just above half.

‘Need’ is not the same as ‘demand’: do ‘in-need’ drinkers want treatment?

All these estimates take no account of whether the drinker wants or intends to reduce their consumption/harm and take a treatment route to achieving their goals. Rather paternalistically, instead they assume that if a research or clinical assessment judges you to be experiencing harm and dependent on alcohol, then you should be in treatment, regardless of how you feel about it. It is, however, entirely conceivable that such drinkers are prepared to shoulder the harm and accept their dependence because they like drinking, or that they want to cut down, but do not see treatment as the solution.

Assessing ‘amenability’ to treatment

Which of the following best describes you?

D I I REALLY want to cut down on drinking alcohol and intend to in the next month.
D I I REALLY want to cut down on drinking alcohol and intend to in the next 3 months.
D I I want to cut down on drinking alcohol and hope to soon.
D I REALLY want to cut down on drinking alcohol but I don’t know when I will.
D I want to cut down on drinking alcohol but haven’t thought about when.
I think I should cut down on drinking alcohol but don’t really want to.
I don’t want to cut down on drinking alcohol.

Fortunately, our key source addressed this issue. Again the method the analysts used had been constructed by them for the Department of Health. The aim was to estimate the proportion of the in-need population who may be “amenable” to treatment – that is, who would enter treatment if it was available and accessible. Their proxy for ‘amenability’ derived from the Alcohol Toolkit Study, which conducts repeated surveys of nationally-representative samples of about 1,700 adults living in private households in England. It offered respondents seven options for describing their desire and intention to “cut down” on their drinking panel. Endorsing any of the first five was interpreted as indicative of desire to cut down (signified by the D logo); if the endorsements included any of the first three, desire was interpreted as accompanied by a concrete intention (signified by the I logo).

Alcohol Toolkit Study surveys also administered the AUDIT screening questionnaire, the same one administered as part of the Adult Psychiatric Morbidity Survey from which our key source had derived its ‘best guess’ ( above) that 595,131 adults in England were in need of treatment. It meant that using AUDIT scores and other variables, results from the two surveys could be cross-referenced to generate an estimate of what proportion judged as in need of treatment by the researchers felt the same way, and might actually enter treatment were it easily available. Based on a desire to reduce drinking at some point, the resulting estimate was that 57% were amenable to treatment, totalling 341,376 alcohol-dependent adults. Narrowing in on those who as well as a desire, endorsed an intention to cut down in the near future, reduced the estimate to 41%, representing 245,614 adults. Respectively these estimates suggest instead of 19%, the treatment caseload in 2014 was equivalent to 33% or 45% of adults who both needed and might consider treatment.

In principle similar calculations could be applied to other estimates of treatment need, including the one ( above) which suggested that in 2014 the treatment caseload was equivalent to 43% of the in-need population. That was based on an AUDIT score of at least 16 allied with the same score on the Severity of Alcohol Dependence Questionnaire, criteria which substantially overlap the criteria used to generate the best-guess 19% estimate. The implication is that the 43% estimate would also be greatly increased by qualifying apparent need for treatment with a desire or intention to cut down indicative of amenability to treatment. What the resulting figure would be is not known, but if the same degree of narrowing applied, the 257,626 in-need population which generated the 43% estimate would be reduced to either 147,877 or 106,400, very close to the estimated treatment caseload of 111,573 in 2014/15. The unknown element is whether the 337,505 people included in the ‘best guess’ estimate who scored as mildly dependent on the Severity of Alcohol Dependence Questionnaire (but a high 20 on AUDIT) would be as amenable to treatment as the narrower bracket which generated the 43% estimate.

Another major source of uncertainty acknowledged by the analysts is whether questions about desire and intention to reduce drinking really are a reasonable indication that the same individuals would consider a treatment route to this objective. We can guess that many would prefer to go it alone or with informal supports, further reducing the numbers both in-need of and amenable to treatment, and correspondingly increasing the extent to which treatment has captured a population equivalent to these numbers. Further uncertainty derives from the absence of the Severity of Alcohol Dependence Questionnaire among the questions used by the Alcohol Toolkit Study, reducing the accuracy with which its findings can be applied to in-need-of-treatment estimates derived from the Adult Psychiatric Morbidity Survey.

Should impaired control be an essential criterion for treatment need?

We might further constrict the population in need of treatment if we accepted that diagnosing an alcohol use disorder requires evidence that, rather than having freely chosen to suffer harm from heavy drinking to gain the perceived benefits, the individual’s ability to control their drinking is pathologically impaired. The resulting ‘harmful dysfunction’ construction is reminiscent of the distinction economists make between harm the drinker risks because for them the benefits are greater still, and harm endured because the grip of addiction impairs the drinker’s ability to act as a rational consumer, weighting up cost against value gained. On balance, harm arising from a free choice may not be counted as a net loss of value for the drinker or for society, while harm due to an addiction-driven non-choice has a greater claim to being considered an uncompensated loss.

Compared to standard clinical criteria for dependence, applying a harmful dysfunction diagnosis to a national US alcohol use survey slashed the numbers calculated as potentially in need of treatment from (according to standard diagnostic criteria) 12.4% in the past year to 2.3%, and the proportion whose supposed need had not yet been met by treatment services from 34% over their lifetimes to just 4%. The markers of dysfunction available from the survey’s questions were imperfect, consisting of experience of adverse effects due to withdrawal symptoms beyond those expected of a ‘hangover’, drinking to avoid these, inability to reduce/stop drinking, and craving for alcohol. The state these signify is that of someone who may want to cut down their drinking and knows they should, but who acts against their will and best interests due to their experience of unwelcome physical or psychological drivers. Applying such a criterion in the UK might well mean that treatment actually captures more patients than the population pathologically driven to suffer harm from repeated heavy drinking.

But there is also a rationale for expanding the in-need population beyond those picked up by questionnaires. Estimates both for England and Scotland based on surveys must be adjusted for the under-reporting indicated by comparing survey responses with how much alcohol is sold. There is reason to believe that the heaviest drinkers underestimate their drinking most when responding to surveys, perhaps really drinking over twice as much as they say. The implication is that the bottom element of the treatment versus need fraction is greater than calculated on the basis of questionnaire responses, amplifying estimates of unmet need.

However, another unknown figure would have a countervailing effect. So far the calculations have accepted that the treatment caseload is the number of patients recorded by the national monitoring system as being in specialist substance use treatment primarily due to their problem drinking (or estimates of these figures derived from the same system). Additional to this number will be some missed patients who should have been in this total, and others who may in some sense be considered to be treated for their drinking problems, but not in specialist services. For many their drinking will be addressed as part of a response to alcohol-related conditions like liver disease or cancer or to alcohol-related crime. Others will be treated in non-specialist settings like primary care or social work and others will be accessing self-help programmes and/or attending mutual aid groups, whose systematic understandings of addiction and of recovery verge on a treatment response, especially when (as in SMART Recovery) groups are led by workers trained in the programme. To the degree that we count these as treatments for problem drinking, to that degree too will the top element of the treatment versus need fraction increase, inflating estimates of the treatment caseload as a proportion of the in-need population, and deflating estimates of unmet need. We can get a sense of the possible order of magnitude of this deflation from the Alcohol Toolkit Study surveys in England during 2014 to 2017. In those samples, 16% who scored as probably dependent on the AUDIT questionnaire (20 or more) and had attempted to cut back on their drinking, said they had attended a specialist alcohol clinic or centre. But just 28% said they had used no supports, meaning 72% may have used some kind of service. The gap between this and the 16% for whom that service was specialist treatment suggests the possible dimensions of the support to reduce drinking not captured by focusing on specialist services. Even if we subtract from the 72% all the 13% who endorsed “Will power/self-discipline”, the gap would remain substantial.

We should also acknowledge an entirely different way of assessing ‘need’ for treatment – not as the population whose drinking, dependence and related problems warrant treatment, but as the population who would experience significant benefits from treatment related to changes in their drinking. This formulation is partly dependent on the nature of the treatment available. The more adequate, comprehensive and individualised it is, the more might benefit.

What is a reasonable target for the proportion of dependent drinkers captured by treatment?

So far we have asked what proportion of the in-need population might actually be in treatment, not what proportion should be in treatment. To answer this question we can draw on specific guidance on what counts for Britain as a good record for engaging in-need drinkers in treatment; less fortunately, its provenance makes it of doubtful validity.

In 2009 the UK Department of Health estimated that provision should be made for 15% of dependent drinkers to access specialist treatment, a figure accepted by NICE. The origin of this figure was an estimate of the demand for treatment in the Canadian province of Ontario, based on a methodology published in 1976 and developed for the US state of Nebraska.

Though perhaps of local applicability, this model does not seem to warrant elevation to an international guide. Its denominator for the population in need of specialised alcohol treatment was derived not from an assessment of harm or dependence, but purely of consumption – the number aged 15 or over who drank at least 475g of alcohol a week, about 59 UK units. The top part of the fraction – the target number for treatment during a year – was not based on an assessment of the proportion of these drinkers who might benefit from treatment, but on the relapse rate (defined as a return to drinking) after treatment and the annual increase in the prevalence of alcohol dependence, in the source study estimated respectively as two-thirds and 10%. To keep pace with relapse of treated patients and the advent of newly dependent drinkers, it was estimated that 15% of the population in need of treatment would have to be treated each year.

Is the path from need to demand to treatment being obstructed?

The upshot of the estimates explored above is that while we may suspect that capturing about 112,000 of England’s problem drinkers in specialist treatment was not enough to meet the need, there is no definitive way to determine whether and the degree to which this was the case. That verdict applies to 2014, but there has been no flowering of new sources of data or new thinking since then which would alter the principles on which it was based, even if the numbers have changed. Worryingly, the way the numbers have changed suggests England is increasingly failing to meet the need for treatment. Public service funding cuts since the age of austerity was initiated in the UK in 2010 are a prime target for the underlying reason why alcohol treatment numbers have been falling despite sustained levels of need.

Alcohol-involved patients in treatment for substance use problems in England from 2005/06

The chart to the right shows numbers falling consistently since 2013/14, whatever criteria is used to identify alcohol treatment. The highest figures show the number of patients whose presenting substance use problems included alcohol, the lowest the number with alcohol as their sole presenting substance use problem. In between are actual or estimated numbers of patients treated primarily for their drinking problems, the basis for the calculations above of the proportion of the in-need population in treatment.

These caseload figures must be married with trends in estimated treatment need to assess whether need is increasingly failing to be met. Using the methodology described above which yielded the 19% treatment versus need estimate for 2014/15, estimates of the alcohol-dependent population in England are available up to 2017/18 and have been back-calculated to 2010/11. In 2017/18 the number was just 1.6% lower than in 2013/14, yet over the same period the estimated caseload of patients treated primarily for their drinking problems fell by about 17% – a rate of decline ten times steeper than that of the estimated population in need of treatment. Calculated on this basis, the treatment caseload fell from being equivalent to just over 19% of the in-need population to 16% chart below right. The drop was within the margin of uncertainty of the population estimates, but it was consistent each year from 2013/14. With a presumed substantial pool of unmet need, even if there had been no fall at all in the proportion of the in-need population being treated, a fall in absolute numbers in treatment would still have been of concern.

Patients in treatment in England primarily for alcohol problems from 2010/11 as proportions of indicators of need for treatment

Another statistic used as an indicator of the need for treatment is hospital admissions of patients diagnosed with mental or behavioural disorders due to drinking, of which harmful use and dependence are the most numerous sub-categories. As with the alcohol-dependent population, the treatment caseload as proportion of admissions has fallen each year since 2013/14, from about 32% that year to 22% in 2018/19 chart right. These trends do indeed suggest that treatment has been capturing smaller and smaller proportions of the in-need population since austerity took hold.

So concerned were Public Health England at the “fall … in the context of high levels of unmet need” that in 2018 they mounted an inquiry, which involved gathering the views of involved parties in nine local authority areas where there had been a fall in numbers, and five where there had been an increase. Its conclusions spotlighted “financial pressures and service reconfiguration”, but also made it clear that the prime service-reconfiguration suspect – the integration of alcohol services with drug services – was itself mainly driven by “reduced local substance misuse budgets”.

All but one of the areas which had seen a fall in numbers had reportedly suffered “budget reductions between 15% and 35%. The wider consultation confirmed this picture and cost savings were often cited as the main reason for service reconfiguration.” Integration of alcohol with drugs was said to have led to a defocus on alcohol and a less specialist response to problem drinking, as well as possibly deterring drinkers from engaging with services which looked and felt like services for drug users.

Also highlighted by respondents were other influences potentially driven by the UK government’s austerity policies, like reduced capacity in wider local health and social care services, the collapse of a national service provider, fewer satellite sites and in-reach sessions in partner agencies and GPs’ surgeries, and pared-down service provision such as replacing individual counselling and therapy with group formats. These influences could be countered by particularly effective local leadership and partnership working, but without these “There is a risk that the falling trend in the numbers in alcohol treatment will continue”.

Rather than declining need as the underlying reason for falling alcohol-treatment numbers, Public Health England’s analysis seems a clear if cautious fingering of austerity-driven public health and other funding cuts; for a well-informed commentator on the response to alcohol in the UK, the time for “treading cautiously” was past: “Some in the field may feel the answer is obvious – continued cuts to treatment budgets (put at 26% for adult and 41% for youth services) have inevitably led to less resources and a changing landscape with very few alcohol-only services remaining, described as a ‘crisis’ in alcohol treatment.”

His views and that of Public Health England’s inquiry were reinforced by a survey of alcohol services and allied professionals in England conducted in 2017 by the campaigning charity Alcohol Change UK. Its key findings were that most respondents could not say there was sufficient local access to these services, and that the main reason was funding squeeze: “In comments, respondents repeatedly said there was simply no money, especially for [rehabilitation] services.” The same year a survey of substance use services in England warned that “the capacity of the sector to respond to further cuts has been seriously eroded”. Instead of targeting the “comprehensive and high quality services” needed to actualise the government’s recovery agenda, service providers were concerned about being able to maintain the basics of “safety and quality in an environment where the pace of change has not yet steadied”.

Another indicator of whether accessible treatment provision is matching demand is the waiting time to start treatment. In respect of drug addiction treatment, good waiting time figures have been seen as showing that treatment supply is keeping up with demand. Good waiting times for alcohol treatment may mean the same; in 2017/18 in England, all but 2% of patients whose sole presenting substance use problem was drinking started treatment within three weeks. However, conceivably this was only because need for treatment was not being converted into demand knocking at the services’ doors, because dependent drinkers were divorced from routes to treatment – much as despite a population being hungry, the demand for bread assessed by visits to bakers could stagnate if the route to the shops is obscure or obstructed – or indeed, if the population does not like the bread they bake.

Alcohol treatment referral sources in England 2008/09 to 2018/19

That this is at least partly the case was suggested by a report on alcohol treatment in England in 2011/12. In theory primary care occupies a pivotal position in identifying need for treatment and converting it into demand by referral to specialist services. But the report expressed concern at how few people had successfully been referred to specialist treatment by GPs or accident and emergency departments, despite the fact that around one in five people seeing a GP is drinking at risky levels, and an estimated 35% of emergency attendances are alcohol-related: “An aim for the coming years is that these two key routes will become more active in identifying and referring people who need treatment for harmful drinking and alcohol dependency”.

If there was cause for concern then, there was even more in subsequent years. Referrals from GPs fell from 14,330 in 2011/12 to bottom at 13,541 the following year, only partially recovering to 13,864 in 2013/14 chart right. From 22%, since 2008/09 the proportion of all treatment entrants accounted for by GPs seems to have fallen each year, reaching 17% in 2013/14. Accident and emergency department numbers and proportions both rose, but from a very low base, peaking in 2013/14 at 1,268 patients, equating to 1.6% of all referrals – still a small proportion of the potential. From a peak of 15,900 in 2009/10, in 2013/14 these two sources accounted for 15,132 treatment starts; as a proportion of all treatment starts, the trend has consistently been down from 23% in 2008/09 to 19% in 2013/14.

From 2014/15 figures for alcohol and drug patients in treatment were merged in the same report. Rather than the alcohol figures representing patients whose primary substance use problem was their drinking, they now were defined as patients who presented with alcohol problems unaccompanied by problems with use of illegal drugs – the ‘alcohol-only’ caseload. However, the trends described above continued chart above. The concurrence between the raw numbers (black lines) and the % these represented of all new referrals (orange lines) shows that the continuing decline was not due to a drop in referrals in general: GPs and accident and emergency departments were not just referring fewer and fewer patients in absolute terms, but also relative to other referral sources. By 2018/19 these sites accounted for just 12% of all new alcohol-only referrals compared to 23% of patients with a primary alcohol problem in 2008/09.

Other UK evidence of treatment missing cases who seem patently in need comes from Wales rather than England, where between 2005 and 2014 just 1 in 4 of the individuals who died from alcohol related causes had been recorded as having at some time been assessed by a specialist substance misuse treatment service.

In the end, Public Health England’s grounds for concern seem a stronger foundation for policy responses than attempts to assess the met-need versus total-need fraction: when there is some hard-to-pin-down but perhaps substantial degree of unmet need, for numbers in treatment to be falling suggests something is increasingly going wrong in access to treatment for problem drinking in England.

Last revised 24 May 2020. First uploaded 24 June 2014

Give us your feedback on the site (two-minute survey)
Open Effectiveness Bank home page
Add your name to the mailing list to be alerted to new studies and other site updates

Top 10 most closely related documents on this site. For more try a subject or free text search

DOCUMENT 2014 How many drinkers should be in treatment?

REVIEW 2010 Alcohol-use disorders: Preventing the development of hazardous and harmful drinking

MATRIX CELL 2020 Alcohol Treatment Matrix cell E2: Treatment systems; Generic and cross-cutting issues

DOCUMENT 2017 Drug misuse and dependence: UK guidelines on clinical management

DOCUMENT 2013 Alcohol treatment in England 2012–13

STUDY 2018 Adult substance misuse statistics from the National Drug Treatment Monitoring System (NDTMS) 1 April 2016 to 31 March 2017

STUDY 2019 Impact of the introduction and withdrawal of financial incentives on the delivery of alcohol screening and brief advice in English primary health care: an interrupted time-series analysis

STUDY 2016 Adult substance misuse statistics from the National Drug Treatment Monitoring System (NDTMS) 1 April 2015 to 31 March 2016

STUDY 2014 Alcohol treatment in England 2013–14

STUDY 2014 Assessing the availability of and need for specialist alcohol treatment services in Scotland